←back to thread

122 points sks147 | 10 comments | | HN request time: 0.212s | source | bottom
Show context
maccard ◴[] No.42165922[source]
As much as I hate it, I think Confluence is the right choice for startups. (Or Github Wiki if you can stomach it). It's free to start, and you get 10 users. By the time you have 10 people who need access, (and you can export PDF's of pages if you have one off sharing to do) you pay $50/mo. If you have 10 employees $50/mo is pocket lint.
replies(3): >>42167177 #>>42168735 #>>42168775 #
1. stavros ◴[] No.42167177[source]
My problem with Confluence is that I want a wiki primarily to be able to find information, and secondarily to store information, and I can never find anything in Confluence. The search function seems to be optimized for maximum frustration.
replies(3): >>42167207 #>>42167327 #>>42168706 #
2. movedx ◴[] No.42167207[source]
Agreed. Overall I find Confluence an anti-pattern and just obnoxious.
3. maccard ◴[] No.42167327[source]
I completely agree with you. I’ve literally searched for page titles and had confluence tell me there’s no results found.

I’ve found this problem exists (even though it’s worse with confluence) no matter what you use, and the solution is a better hierarchy

replies(1): >>42167397 #
4. stavros ◴[] No.42167397[source]
I cannot for a minute believe that search over a few hundreds/thousands of documents is anything less than a solved problem these days. Confluence is another level of bad, but I'm sure some software does search well.
replies(1): >>42171559 #
5. 9x39 ◴[] No.42168706[source]
That's really surprising. Any idea if it was Confluence self-hosted vs cloud?

We have a pretty substantial Confluence Cloud and don't see the "search doesn't work" problem. Tangent: we have people who browse, and people who search, and ne'er the twain shall meet, but that's a different problem in page structure, conventions, tags, whatever.

replies(1): >>42169033 #
6. stavros ◴[] No.42169033[source]
It's Confluence cloud. I've looked for documents I myself have written, using the keywords I knew were in the title, and got irrelevant random pages. It's ridiculous.
replies(1): >>42169247 #
7. Arrowmaster ◴[] No.42169247{3}[source]
This was my experience as well except even worse. When my employer was acquired by evil mega corp ISP, they took away our mediawiki and replaced it with confluence. But they also didn't give us write access and instead had a new technical documentation team doing it except they were just uploading Word docs to Confluence. We couldn't search for anything in the docs but even searching for article titles failed so we all just had to keep bookmarks to what we needed.
replies(1): >>42171542 #
8. maccard ◴[] No.42171542{4}[source]
> We couldn't search for anything in the docs but even searching for article titles failed so we all just had to keep bookmarks to what we needed.

My experience with _every_ knowledge base has been that this is the solution. My current team has confluence pages pinned to slack channels which link to the other necessary pages.

9. maccard ◴[] No.42171559{3}[source]
I agree. If google can index the entire internet, there's no reason confluence can't index 10 pages by title in the length of time it takes to use the search.

But, my experience is that even when search is good, you need the right query. Take onboarding instructions. Do you search for "new hire", "initial setup", "setup guide", "engineering onboarding", "programmer first steps"?

Or, do you send a link in a welcome email to the right page, and have it pinned on the home page of your project?

Honestly, I think confluence would be better _without_ search.

replies(1): >>42172128 #
10. stavros ◴[] No.42172128{4}[source]
I don't always want to find a page in response to something else. If I'm looking for a decision document I wrote a year ago, how am I going to do that without search?